You can deploy this model and start earning money today!

Views

No views yet

🏥 TANIT-Med-8B v1.0 (Final)

Name: TANIT-Med-8B-v1
Availability: InStock

Advanced Medical Reasoning Model by TANIT Healthcare Technologies

License

Model

Phase

🌟 Model Overview

TANIT-Med-8B v1.0 (Final) is a specialized medical reasoning model developed by TANIT Healthcare Technologies.

Production-ready medical reasoning model. Fine-tuned through 4-phase training: Broad SFT → Deep Reasoning → DPO Alignment → Ultra-Refined Polish.

Key Features

🧠 Advanced Medical Reasoning: Deep chain-of-thought reasoning for complex medical scenarios
🔬 Evidence-Based Responses: Trained on validated medical knowledge sources
💭 Transparent Thinking: Exposes reasoning process via <think> tags
⚡ Efficient: 8B parameters optimized for deployment
🛡️ Safety-Aligned: DPO-trained for safer, more helpful responses

🔬 Training Pipeline

TANIT-Med-8B was developed through a carefully designed 4-phase training pipeline:

┌─────────────────────────────────────────────────────────────────┐
│  Phase 1: Broad Medical SFT                                     │
│  ├── 150K diverse medical samples                               │
│  ├── Foundation knowledge acquisition                           │
│  └── Loss: 1.129 → Comprehensive medical understanding          │
├─────────────────────────────────────────────────────────────────┤
│  Phase 2: Deep Reasoning SFT                                    │
│  ├── 50K high-quality reasoning chains                          │
│  ├── Focus: Medical-R1-Distill + Huatuo-o1                      │
│  └── Loss: 1.014 → Enhanced reasoning capabilities              │
├─────────────────────────────────────────────────────────────────┤
│  Phase 3A: DPO Preference Alignment                             │
│  ├── 33K preference pairs (FineMed-DPO)                         │
│  ├── Direct Preference Optimization                             │
│  └── Loss: 0.26 → Human-aligned responses                       │
├─────────────────────────────────────────────────────────────────┤
│  Phase 3B: Ultra-Refined Polish                                 │
│  ├── 22K DeepSeek-R1 distilled samples                          │
│  ├── Ultra-low LR (3e-7) to preserve alignment                  │
│  └── Loss: 0.828 → Production-ready model                       │
└─────────────────────────────────────────────────────────────────┘

🚀 Quick Start

Installation

pip install transformers torch accelerate

Using with Transformers

python

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_name = "TanitAI/TANIT-Med-8B-v1"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True
)

# Medical question
question = """A 45-year-old male presents with sudden onset chest pain radiating to 
the left arm, diaphoresis, and shortness of breath. ECG shows ST elevation in 
leads V1-V4. What is the most likely diagnosis and immediate management?"""

messages = [
    {"role": "user", "content": question}
]

# Apply chat template
input_text = tokenizer.apply_chat_template(
    messages, 
    tokenize=False, 
    add_generation_prompt=True
)

inputs = tokenizer(input_text, return_tensors="pt").to(model.device)

# Generate response
with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_new_tokens=2048,
        temperature=0.6,
        top_p=0.95,
        do_sample=True
    )

response = tokenizer.decode(outputs[0][inputs['input_ids'].shape[1]:], skip_special_tokens=True)
print(response)

Using with vLLM (Recommended for Production)

python

from vllm import LLM, SamplingParams

model_name = "TanitAI/TANIT-Med-8B-v1"

llm = LLM(
    model=model_name,
    dtype="bfloat16",
    tensor_parallel_size=1,  # Adjust based on your GPU setup
    trust_remote_code=True,
    max_model_len=8192
)

sampling_params = SamplingParams(
    temperature=0.6,
    top_p=0.95,
    max_tokens=2048
)

question = "What are the diagnostic criteria for Type 2 Diabetes Mellitus?"

# Format with chat template
prompt = f"<|im_start|>user\n{question}<|im_end|>\n<|im_start|>assistant\n"

outputs = llm.generate([prompt], sampling_params)
print(outputs[0].outputs[0].text)

Using via API (Without Loading Model)

For team members who need access without loading the model locally, use the HuggingFace Inference API:

python

import requests

API_URL = "https://api-inference.huggingface.co/models/TanitAI/TANIT-Med-8B-v1"
headers = {"Authorization": "Bearer YOUR_HF_TOKEN"}

def query(payload):
    response = requests.post(API_URL, headers=headers, json=payload)
    return response.json()

# Example query
output = query({
    "inputs": "What is the first-line treatment for hypertension?",
    "parameters": {
        "max_new_tokens": 1024,
        "temperature": 0.6
    }
})
print(output)

Accessing via HuggingFace Hub

python

from huggingface_hub import InferenceClient

client = InferenceClient(
    model="TanitAI/TANIT-Med-8B-v1",
    token="YOUR_HF_TOKEN"
)

response = client.text_generation(
    "Explain the pathophysiology of heart failure.",
    max_new_tokens=1024,
    temperature=0.6
)
print(response)

📊 Evaluation Results

Benchmark	Test	Test Hard
Average	50.7%	20.3%
MedQA	65.0%	32.0%
PubMedQA	68.0%	21.0%
MedMCQA	57.8%	32.0%
MMLU (Medical)	77.5%	26.0%
MMLU-Pro (Health)	47.3%	22.0%
MedExQA	73.3%	15.0%
MedBullets	42.9%	15.7%
AfriMedQA	48.9%	18.8%
MedXpertQA-R	12.1%	11.0%
MedXpertQA-U	14.6%	9.0%

Evaluated using zero-shot prompting with chain-of-thought reasoning.

📋 Training Details

Parameter	Value
Base Model	`deepseek-ai/DeepSeek-R1-0528-Qwen3-8B`
Training Phase	Final (v1.0)
Training Data	22K high-quality samples from Medical-R1-Distill (DeepSeek-R1 distillation)
Training Steps	687
Training Time	0.8 hours
Final Loss	0.828
Precision	bfloat16
Context Length	8,192 tokens

⚠️ Limitations & Intended Use

Intended Use

Medical education and research
Clinical decision support (with physician oversight)
Medical question answering
Healthcare documentation assistance

Limitations

Not a replacement for professional medical advice
May generate plausible-sounding but incorrect information
Performance varies across medical specialties
Should always be used with human oversight in clinical settings

Ethical Considerations

This model is intended to assist, not replace, healthcare professionals
Always verify medical information with authoritative sources
Do not use for diagnosis or treatment without professional consultation

📜 License

This model is released under the Apache 2.0 License.

🙏 Acknowledgments

Built on DeepSeek-R1-0528-Qwen3-8B
Training data from FreedomIntelligence, UCSC-VLAA, Intelligent-Internet, and other open-source medical datasets
Developed by TANIT Healthcare Technologies

📧 Contact

For questions, collaborations, or access requests:

Organization: TANIT Healthcare Technologies
Email: contact@tanitai.com

Made with ❤️ by TANIT Healthcare Technologies

Advancing healthcare through AI